Understanding Value at Risk (VaR)

Value at Risk estimates the maximum expected loss over a chosen horizon at a chosen confidence level.

Every VaR figure depends on three inputs: confidence level, time horizon, and the portfolio being measured.

The three core VaR methods are parametric, historical simulation, and Monte Carlo simulation.

VaR is useful, but it says nothing about the size of losses once the tail threshold has been breached.

Expected Shortfall is a stronger tail-risk measure and has replaced VaR in key regulatory frameworks like FRTB.